A spelling error was the tipoff to last month’s multimillion-dollar digital bank heist. But could multifactor authentication have prevented it in the first place?
The frightening prospect of a coordinated cyberattack on the global financial system has been the subject of seemingly endless speculation and the stuff of fictional novels. Last month, a group of international cybercriminals provided resounding proof that a multimillion-dollar digital heist is no longer a hypothetical scenario.
On February 4 and 5, a group of cybercriminals using system credentials reportedly stolen from the Central Bank of Bangladesh, transferred $81 million from an account at the New York Fed to various accounts in the Philippines. The electronic thievery was carried out through four wire transfers using the SWIFT payment network, an internationally trusted medium primarily used for issuing institutional payment instructions. Had it not been for an alert Deutsche Bank employee, the routing bank on the fifth transaction, the total take of this heist may have been upwards of $1 billion.
The employee noticed a misspelled word on the fraudulent request, causing Deutsche Bank management to seek additional clarification. The $20 million transaction that ultimately revealed the scheme was intended for an obscure Sri Lankan nonprofit, the Shalika Foundation. Only after the conscientious employee noticed that the wire request read “Fandation” rather than “Foundation,” did authorities begin to unravel the plot and identify dozens of similar requests awaiting processing at the New York bank.
That’s right, this brazen scheme went unnoticed by both central banks and was ultimately detected not by the institutions’ legion of network security professionals or advanced technological threat indicators, but rather by an employee that noticed a spelling error. Are you concerned about your savings accounts and retirement plans yet?
This incident has rightly caused concern throughout the financial services industry and has stoked the fears of those who rely on the critical process by which electronic monetary transfers are made. While international law enforcement and intelligence agencies are collaborating to identify those responsible for this operation, Bangladeshi and American officials are publicly posturing.
The Bangladeshis maintain that the American central bank should have noticed that one of the requests was directed to an unregistered Sri Lankan charity, thereby triggering a fraud alert and additional investigation. They further maintain that the New York bank should have immediately viewed this transfer request with suspicion based on the fact that it was not intended for another bank and that a transfer had never before been made to this organization.
The media has reported that the Bangladeshi SWIFT account login credentials were obtained from a keylogger that had been surreptitiously installed within the bank’s network several weeks before the heist. Although the United States has been restrained in its comments regarding the matter, a Fed spokesman has stated “there is no evidence that Fed systems were compromised.” In other words, “don’t blame us if you can’t secure the access credentials for your most sensitive financial systems.”
Both sides are right.
Without question, the Fed should have been capable of detecting that part of the money was destined not to another financial institution, but to a recently established, foreign entity. Regardless of how the cybercriminals gained access to the Bangladeshi SWIFT login credentials, an automated alert mechanism should have been in place to alert the Fed that the intended recipient was not a known financial institution and that transfers had never before been made to this account. These are characteristics that are present in many institutional cyber fraud campaigns and should have been detected. If a retail bank is capable of implementing an automated notification process when a consumer’s credit card is used under suspect circumstances, then the world’s most influential central bank should be able to identify and disrupt a suspicious, multimillion-dollar wire transfer in real-time. The bells should have been ringing and the warning lights flashing.
Alternatively, the Bank of Bangladesh should have better safeguarded the system credentials used to facilitate this theft. If, in fact, the SWIFT account’s login credentials were obtained by a keylogger, then it is highly unlikely that multifactor authentication was in place to protect this highly sensitive account. Multifactor authentication would have required the user to possess at least one dynamic identifier to gain access to the account, thereby neutralizing the credentials captured by a keylogger because one of the passwords would change at each session login. If multifactor authentication was not in place, this would constitute a catastrophic failure of basic access control for a process as sensitive as this.
This incident represents an inexcusable, collective failure of basic security protocol and has confirmed the long held fear that the world’s central banks are subject to well-coordinated cyber campaigns. Given that these institutions represent the foundation of global commerce, it is critical that those responsible for securing the data and monies held therein remain vigilant.
John Moynihan is President of Minuteman Governance, a Massachusetts cybersecurity consultancy that provides services to public and private sector clients throughout the United States. Prior to founding this firm, he was CISO at the Massachusetts Department of Revenue from … View Full Bio